ARROW-14311: [C++] Make GCS FileSystem tests faster

[coryan]

Avoid using the default credentials for the unit test. In some
environments these can be very slow to initialize, as they test several
different potential sources of credentials, some including HTTP
requests.

[github-actions]

https://issues.apache.org/jira/browse/ARROW-14311

[coryan]

@pitrou would you be able to test this in your environment? I have good reason to believe (strace output) that this fixes the problem. An independent check would be helpful, as it was working well in my environment.

[pitrou]

It works indeed :-)

[==========] Running 5 tests from 2 test suites.
[----------] Global test environment set-up.
[----------] 4 tests from GcsFileSystem
[ RUN      ] GcsFileSystem.OptionsCompare
[       OK ] GcsFileSystem.OptionsCompare (0 ms)
[ RUN      ] GcsFileSystem.ToArrowStatusOK
[       OK ] GcsFileSystem.ToArrowStatusOK (0 ms)
[ RUN      ] GcsFileSystem.ToArrowStatus
[       OK ] GcsFileSystem.ToArrowStatus (0 ms)
[ RUN      ] GcsFileSystem.FileSystemCompare
[       OK ] GcsFileSystem.FileSystemCompare (2 ms)
[----------] 4 tests from GcsFileSystem (2 ms total)

[----------] 1 test from GcsIntegrationTest
[ RUN      ] GcsIntegrationTest.MakeBucket
INFO:werkzeug: * Running on http://localhost:60709/ (Press CTRL+C to quit)
INFO:werkzeug: * Restarting with stat
INFO:werkzeug:127.0.0.1 - - [15/Oct/2021 19:00:25] "POST /storage/v1/b?project=ignored-by-testbench HTTP/1.1" 200 -
INFO:werkzeug:127.0.0.1 - - [15/Oct/2021 19:00:25] "GET /storage/v1/b/test-bucket-name HTTP/1.1" 200 -
[       OK ] GcsIntegrationTest.MakeBucket (1551 ms)
[----------] 1 test from GcsIntegrationTest (1551 ms total)

[----------] Global test environment tear-down
[==========] 5 tests from 2 test suites ran. (1554 ms total)
[  PASSED  ] 5 tests.

[pitrou]

(1.5s make still be a bit long once there are a lot of tests, but perhaps it's not really possible to make it faster)

[coryan]

(1.5s make still be a bit long once there are a lot of tests, but perhaps it's not really possible to make it faster)

I will see what I tolerate. At some point I might start the testbench once, and then isolate the tests using different buckets or something similar.

[coryan]

Is there something else I need to do here or can we merge this PR?

[coryan]

Ping. Is there are reason we cannot merge this PR?

[kou]

I'll merge this.

@pitrou If you have any concern, please write it here.

[ursabot]

Benchmark runs are scheduled for baseline = 1bac505 and contender = 1b7178d. 1b7178d is a master commit associated with this PR. Results will be available as each benchmark for each run completes.
Conbench compare runs links:
[Finished ⬇️0.0% ⬆️0.0%] ec2-t3-xlarge-us-east-2
[Finished ⬇️0.51% ⬆️0.0%] ursa-i9-9960x
[Finished ⬇️0.13% ⬆️0.18%] ursa-thinkcentre-m75q
Supported benchmarks:
ursa-i9-9960x: langs = Python, R, JavaScript
ursa-thinkcentre-m75q: langs = C++, Java
ec2-t3-xlarge-us-east-2: cloud = True

[pitrou]

@coryan Sorry for the delay. I was on holiday.